We shall discuss several issues: the nature of a document, and ways of classifying it; conceptual and practical matters of documentary research; ways of analysing documents; and finally some critical remarks on documentary research.

Documentary research method has had little attention compared to other methods because of the dominance of positivism and empiricism so that statistics and quantification are popular forms of data collection and analysis. Also, documentary research is associated with historical research, and history sits uneasily alongside social science disciplines. Documentary research is regarded as being not clear-cut, not having a method and nothing on how a researcher uses it. Yet these criticisms are misplaced. History as a discipline provides us with a sense of our past and with that the ways in which our present came about, and employing a range of documentary sources (e.g., state government files) enables researchers to reflect on contemporary issues.

Researchers can also see how new questions arising from contemporary concerns (e.g., gender issues, racism and nationalism) may turn them back towards an interrogation of history to uncover those aspects of the past that had remained hidden from view.

A broad definition of a document is a written text. Writing is the making of symbols representing words and involves the use of a pen, printing machine and other tools for inscribing messages on some material medium (such as paper). Similarly, electronic means of storing and displaying texts (files and documents) are also true documents. Documents may be regarded as physically embodied texts, where the containment of the text is the primary purpose of the physical medium. Sources of documentary research include historical documents such as laws, declarations, statutes and people’s accounts of events and periods. Also, reports based on official statistics would be covered, as well as governmental records, mass media, novels, plays, drawings, and personal documents such as dairies and biographies. To this list, photographs can be added – though they lie between aesthetic and documentary record.

There are three ways of classifying documents.

• primary, secondary and tertiary documents: Primary sources refer to those materials which are written or collected by those who actually witnessed events which they describe. While considered to be reliable and accurate, such sources have to be seen in their social context (i.e., selective and bias reporting), and for this reason the researcher may employ secondary sources. These are written after an event that the author has not personally witnessed, and the researcher must be aware of problems (i.e., incomplete and prejudice coverage) in the production of this data. Tertiary sources enable researchers to locate other references such as indexes, abstracts and other bibliographies. Libraries are sites for such collections of tertiary sources. Internet search engines can also be added.
• public and private documents: Documents can divided into four categories according to the degree of their accessibility: closed (e.g., secret police files), restricted (e.g., medical files and confidential corporate reports), open-archival (e.g., census reports) and open-published (e.g., government budget statistics).
• solicited and unsolicited documents: some documents (e.g., government surveys and research projects) would have been produced with the aim of research in mind, whereas others (e.g., diaries) would have been produced for personal use. However, even if unsolicited documents are for personal use, they still are addressed to an audience.

There are several ways in which researchers might conceptualise a document and frame their research questions accordingly.

According to positivists, documents (e.g., news reports and crime statistics) represent an objective reflection of reality. They become mediums through which the researchers search for a correspondence between their descriptions and the events to which they refer. Here, documents account for social facts that exist independently of interpretations.

In contrast, for interpretivists (in particular, ethnomethodologists), documents (indeed, news and crime reports) are representatives of practical requirements and constructed for those purposes. Here, documents are seen as ‘accomplishments’ that construct what happened, based on a form of practical reasoning that renders the social order accountable and comprehensible, and yet also open to negotiation and manipulation by interested parties. Documents are to be used to examine unstated, tacit and implicit meanings and structures embedded within the documents, as they refer to some underlying social patterns or values. Like interviews and statistics, documents are seen as topics, not as resources – i.e., to be used to make sense of how social actors construct reality, rather than to be used to reflect it.

For critical realists, documents cannot be read in a detached manner. Instead, researchers must approach documents in an engaged, not detached, fashion. The analysts must consider the differences between their own frames of meanings and those found in the text. A researcher begins with an analysis of the common sense procedures that came to formulate the document (for instance, short term cost saving measures generating part-time and temporary employment contracts), and then locates it within a wider social and political context (e.g., free market forces and weakening collective bargaining and trade union powers). Researchers next examine the factors surrounding the processes of its production and social context (say, neo-liberalism and marketisation of society). It then becomes necessary to rise above the particularity of the text and the rules of the context to produce an understanding and explanation of the text in the context of wider political, economic and social forces. Here, documents are viewed as media through which social mechanisms, structures and powers are expressed. They are approached in terms of the cultural and economic context in which they were written, and may be viewed as attempts at persuasion. Approaching documents in this way tells researchers a great deal about the societies in which writers write and readers read. A document (e.g., a newspaper) might, for example, reflect the marginalisation of particular groups of people and the social characterisation of others (say, social deviant groups such as gays and lesbians).

For feminists, a text (e.g., an advertisement featuring scantily-dressed women) is a reading to revise its premises and also an act of refusal. The researcher concentrates on the way in which the text constructed the contribution of women to an event, but the strategy of refusal enables women to see their contemporary social and political situation in a new light (i.e., women exploitation and oppression). Here, feminist researchers, and women in general, must understand the assumptions behind the text, and resist, refuse and revise them to enable themselves and women to liberate themselves from patriarchal powers.

Critical theorists share much of critical realists’ approach to texts. They focus not only on the relationship between the author and the document, but the ways in which the use of a document (say, a mythical story) is linked to the present as acts of historical writings are inevitably linked to current uses (say, nationalism and identity).

The post-structuralist approaches suggest that a text (e.g., media) does not refer to anything beyond itself nor to the intentions of its author. Here, the debate whether the text is (which would include an interview transcript or observation field notes) a topic of social research, or a resource for social research is revived.

Generally, theorists suggest that a researcher should approach a document in terms of three levels of meaning and interpretation:

• the meanings that the author intended to produce;
• the received meanings as constructed by the audience in differing social situations;
• the internal meanings that semioticians exclusively concentrate upon.

There are often practical impediments to the research aims of collecting and analysing documents:

• lack of time, finance and resources to undertake a full inquiry;
• the relevant data may not be available;
• gatekeepers to the sources may not grant access to it;
• the publication of materials (e.g., criminal and sexual affairs) may be detrimental (and so unethical) to the memory of an author or group of people;
• the publication may lead to a public outcry (and so politically insensitive) that would not advance understanding of a process or event;
• the data (especially hand-written documents) may be difficult to read;
• the historical and local documents may have been damaged;
• the documents (especially personal sources) may use abbreviations and coded references for individuals, practices and relations that may be difficult to interpret;
• the documents are always context-specific, so making empathy and understanding quite demanding.

In such cases, imagination, along with an understanding of the issues and methods of social research, is often required to gather data from a wide range of documentary sources (i.e., primary, secondary and tertiary sources, private and public documents and solicited and unsolicited reports).

There are four criteria for assessing the quality of the evidence available from documentary sources.

• Authenticity: There are several guidelines for assessing a document’s authenticity:

• the document may contain obvious errors, or is not consistent in its representation;
• different versions of the same document exist;
• there are internal inconsistencies in terms of style, content and handwriting, and so on;
• the document has been in the hands of a person or persons with a vested interest in a particular reading of its content;
• the version derives from a suspect secondary source;
• it is inconsistent in relation to other similar documents;
• it is ‘too neat’ in terms of being representative of a certain group of documents.

• Credibility: This refers to the extent to which the evidence is undistorted and sincere, free from error and evasion. Are the people who record the information reliable in their translations of the information that they receive? How accurate were their observations and records? To achieve this, researchers may employ other sources, and question the political sympathies of the authors.
• Representativeness: The issue of whether a document is typical depends on the aim of the research. If the researchers are concerned with drawing conclusions which are intended to argue that there is a ‘typical document’ or ‘typical method’ of representing a topic in which they are interested, then this is an important consideration in order to demonstrate how one interpretation of an event predominates to the exclusion of others. However, ‘untypical’ documents may be of interest.
• Meaning: This refers to the clarity and comprehensibility of a document. While meanings change and the use of words varies, social context enables understandings. Here, meaning can be divided into intended, received and content/internal meaning.

Documents do not stand on their own, but need to be situated within a theoretical frame of reference in order that its content is understood. For this purpose, we can use content analysis, which compromise three stages: stating the research problem, retrieving the text and employing sampling methods, and interpretation and analysis. Content analysis takes both quantitative and qualitative forms.

It seeks to show patterns of regularities in content through repetition. The analyst would seek to derive categories from the data (e.g., a daily newspaper) in order that it can be compared with other data. Words and phrases (say, ‘truth’, ‘liberties’, ‘freedom’, ‘opposition parties’, ‘civic associations’, and so on) in the documents are transformed into numbers. The number of times in which a word or phrase occurs in the text is taken as an indicator of its significance in the text (e.g., more references to ‘freedom’ than ‘authoritarianism’ may show the political tendencies of the editor and owner of the newspaper). This quantification strategy is assumed to enhance both the reliability and validity of the classified data. It is taken for granted that there exists a defensible correspondence between the transformed account and the way the information was meant in its original form.

However, there are several weaknesses with this quantification approach.

• quantification deals only with what has been produced (i.e., the frequency of words and phrases), not with the decisions that informed its production, which tells the researcher so much more about the its intended and received meanings;
• it only deals with information that can be measured and standardised, and therefore provides only a simplified system of categories of data;
• it tends to reproduce the meanings used by authors as opposed to subjecting them to critical analysis in terms of the political, social and economic context of their production;
• it fails to understand the common-sense context in which the words and phrases are produced and interpreted as people make sense of their social world;
• it is assumed that the audience/reader who receives the text must translate it as the researcher does.

In short, the approach negates the idea that a text is open to multiple and contradictory readings by its audience. It is wrong to assume that the textual interpretation can be simply read off without considering the readers’ understanding. The frequency with which words or phrases occur in a text may therefore say nothing about its significance within the document.

For researchers to grasp a document’s significance, they must concentrate on intended, received, and content/internal meanings. Qualitative analysis views the author as a self-conscious actor addressing an audience under particular social and political circumstances. The task of the researcher is to read the text in terms of its symbols – as an anthropologist does with rituals. This ‘reading’ may be derived from secondary sources and/ or other research methods such as observational studies. The researcher considers not only how existing interpretations are constructed, but also how new ones are developed and employed. While author’s intended meanings are important, analysing the reader’s social situation is also crucial to interpreting the text.

Aside from the emphasis on intended and received meanings, there is also content meaning upon which content analysts and semioticians focus their attention. This examines the relationship between a signifier (i.e., a symbol or word) and a signified (i.e., a concept or idea to which the signifier refers to), and its relationship to a referent (i.e., a material object or language system). The signified may not refer to a material object, but instead refer to the way in which a system of language (through its signs) organises the world. Content analysts focus upon the relationships within the text, and its relationship to other texts. A critical-analytic stance considers how the document (e.g., a government news report) represents the events that it describes and closes off potential contrary interpretations (e.g., oppositional understandings) to the reader. This stance considers the way in which a text attempts to stamp its political, cultural and economic authority upon the social world it describes. In so doing, the social world might be characterised by the exclusion of valuable information (e.g., on women and minority groups) and the characterisation of events and people in particular ways according to certain powerful interests.

There are several criticisms of documentary research methods.

• Given its social context and identity, the researcher will give a selective and biased understanding of a document, and may even deliberately chose and select particular documents. This is a common criticism against researchers using any technique of data collection.
• Inevitably, authors of documents will decide to record and leave out information informed by their social, political and economic environment of which they are part. Historical documents, thus, are amenable to manipulation and selective influence. In undertaking documentary research, researchers should be aware of these influences and not assume that documents are simply neutral artefacts from the past. Indeed, uncritical readings of texts can reproduce and reinforce marginalisation of groups, such as working class, women and ethnicities.
• While new technologies (e.g., the internet) offer possibilities for acquiring documents, researchers have to exercise a critical reflexivity since much of the documents on the internet are produced by powerful political, cultural and economic groups, who want to ensure that particular images reach the public domain, and wish to counter bad images with more favourable representations.
• While digital technology (e.g., computers) provides opportunities for new forms of research, it should be remembered that it lends to the creation, modification, destruction and replacement of information with very little effort and cost.
• More generally, using documents without due consideration to the process and social context of their construction, leaves researchers (especially semioticians) open to the charge of being unreflexive and uncritical in their ‘readings’.